2700x 300mhz faster than 1700x

2700x 300mhz faster than 1700x

just like 2600 is only 300mhz faster than 1600

>yfw amd is making kaby lake look like a good skylake upgrade

videocardz.com/75159/amd-ryzen-7-2700x-spotted-with-clocks-up-to-4200-mhz

Other urls found in this thread:

techpowerup.com/241593/amd-ryzen-2000-series-pinnacle-ridge-cpus-get-soldered-ihs
en.wikipedia.org/wiki/Bubble_(computing)
hothardware.com/news/amd-confirms-am4-socket-support-future-ryzen-processors-2020
twitter.com/NSFWRedditVideo

╲⎝\╲⎝\╲⎝\AMD BTFO⧸⎠╱⧸⎠╱⧸⎠╱

as expected from AMD

No they aren't, Kabylake offered literally fucking nothing over Skylake.

At least you don't need to pay extra to overclock them.

Isn't the 2000 series Ryzens just Zen+ architechture though? They're pretty much literally just a rehash of the first gen with better optimization

It's a tweaked Zen architecture, mostly to do with improved turbo algorithms and reduced memory latency.
That still nets a IPC improvement, more than any single generational IPC improvement from post Haswell.

That's an OC on all cores though so performance will be better. We still don't know what the true OC potential will be. The 2800x should be around 4.3Ghz on alll cores. Also clock speed is not everything. Other optimizations and how long it can hold that clock speed are another factor. Intel's clock speeds generally fall back after a while.

HAHAHAH HAHAHAH HA FUCK U AMD SHILLS GO BACK TO PLEBBIT

I hate these retarded graphs that don't start from 0 to make everything look bigger than it is.

>Same TDP
Not sure if we'd even see gains if they've improved their process considerably. OCing might see extra headroom still.

Numbers are included, only someone brain damaged would misinterpret the data in that image.

Put your tripcode back on samefag

I seriously doubt the 2600 was turboing all cores to 3800 in those tests, the MT score difference might be larger.

Cry more faggot

>phoneposter
>edited image

>having to prove it to add legitimacy to shitposting

That is next level

Okay head on back now children

fuck off faggot
i don't use reddit

What about Nvidia?

Exactly what a plebbitor would say

Less than Intel, Intel is much bigger and AMD has billion of fanboys

Seems like an ok little improvement. Looking forward to seeing if they will overclock better.

How much is out about the new chipsets?

Lying doesn't work when the data you're lying about is easily acquirable.

Dumb Nvidiot.

Nothing besides changing the PCIe2 lanes to PCIe3

There's no lack of SATA/USB support already.

then what do you expect me to say?
i hate that shitty platform

MUH GIGAHURRRRZ!!1

>mhz
That doesn't matter as much as you seem to think. It says close to nothing about real world performance. ...not that it doesn't give some indicator as to how these chips will perform. My guess is around 3-5% better.

I've said that the AM4 "upgrade path" "selling point" bullshit shills here has been silly and stupid numerous times before. It should be more obvious now that the chips are here. You don't buy a $300 1700x just to "upgrade" to a max 5% faster $300 2700x a year later.

Yes. It's the same shit in a new wrapping with some minor adjustments that allow for slightly higher clocks and perhaps slightly higher IPC.

It's pretty amazing how much people will bend over for 500MHz

Anyone tried Project Lasso?

>You don't buy a $300 1700x just to "upgrade" to a max 5% faster $300 2700x a year later
Speak for yourself. I'm personally "computer guy" for two families where exactly this is happening.
One where the father is getting a cheap motehrboard and the son's 1700X and one where the son is getting the father's 1800X

>You don't buy a $300 1700x just to "upgrade" to a max 5% faster $300 2700x a year later.
That's why you wait for 7nm Zen2 instead.

...

Stop shilling juan

looks like im getting coffeelake

That's not what redditors say, redditors say "Hating reddit is a meme, most people use both websites"

>amd is making kaby lake look like a good skylake upgrade
Meanwhile here in the real world, Kaby Lake was 200MHz faster than Skylake

>4.2 stock oc
>not hitting a wall at 4ghz
seems pretty okay

>essentially a 4.2ghz 1700x with a little IPC boost
>also this is just the refresh
Seems like your wojak should be the crying intel version because this is going to be great

>a simple refresh is 3.68% better in multicore and 8.64% better in single core
Intel on literal suicide watch

Stupid meme. K series processors have a higher base clock, so you're not just paying for OC ability.

>using meltdowntel in 2018

I will never use that jew trash again.

Nvidia related subreddits tend to be either wastelands or focused on very specific things like GPGPU, etc.

>mhz

Haven't you intel cucklords learned your lesson and gotten over your marketing brainfarts?

>muh mhz

if a 2700 can hit 4.5ghz with stock cooler on a B350 motherboard then I will rebuild my system around it

>Globalfoundires say 10-15%
>AMD says 10-15%
>13.3% higher clockspeed
>Apparently this is bad
13.3% clockspeed improvement coupled with 3.6% single thread IPC improvement is ~18% single thread performance increase.
Better than Intel, yes.

.3 GHz is a fucking great bump you faggot

>tweak something to improve memory latency
>foundry refresh gives clock speed bump
>get solid performance increase at ISO power without any arch changes
>people complain

Why

tryingwaytoohard/10

Without meltdown overhead as well, which I doubt intel benchmarks account for.

>just like 2600 is only 300mhz faster than 1600
>link clearly shows 2600 is 200 MHz faster than 1600
It's already boosting as high as the best OG Ryzen overclocks out of the box, and the 2700x isn't even the top SKU.

Of course you don't, but upgrading, say, from a 1600 to a 3700 would be a substantial boost

that isnt fair cpus arent allowed to improve process and ipc at the same time fucking amd cheats. bench SYSmark instead

stock cooler will never happen

If they OC to at least 4.8 there is literally no reason to buy Intel.

>What is a refresh

The real upgrade will come with Zen2. You are fool if you "upgrade" from OG Zen.

I'm gonna do it if it can overclock to 4.4. Maybe even if it can overclock to 4.3, since my current chip can't do better than 3.9.

kek

JUST W*IT ™

...

This. It'll be interesting to see how compatible Zen2 will be with X370.

This is great 300MHz plus better IPC, I was only expecting a 200MHz increase.

Intel BTFO once again.

2700X already does 4.3 with XFR, 2800X is very likely to hit 4500 and threadripper might get another 100MHz.

A 400MHz advantage is all Intel will have, that's after deliding and wasting $80 on cooling since their piece of shit 4 core can draw close to 200W overclocked.

that's only a 300mhz boost over stock clocks, faggot

Man AMD are on a roll, those Raven Ridge APUs are real nice little systems give them a couple of months for the software optimization to shake out and there will be a solid platform.

Now these Zen+ seeing more gainz for the heavier user, and the quality laptop components out there AMD are looking good. Has Intel even patched Meltdown yet? I can't even tell anymore.

>Intel even patched Meltdown yet? I can't even tell anymore.

Meltdown was patched eons ago. Spectre V2 are still not patched for Xeon E7s and Pre-Haswell CPUs

>You don't buy a $300 1700x just to "upgrade" to a max 5% faster $300 2700x a year later.
but isn't that what intel has been doing since sandy? only they make you pay to overclock and jew you on motherboards. oh and jizz TIM.

No it's not. :^)

Source: your :^)

It really is okay when Intel does it. Lets take kaby lake for example - when it released nobody cared as it was just hot clocked skylake with a socket change. As soon as zen released kaby lake was a game changer of an architecture, pinnacle of Intel's might.

Finally some nice IPC improvements over

MOAR NIGGAHURTZ LMAO

I know this sounds strange to an Intel user, but motherboards can in fact be used for processors more than a single year apart.

Good goy

FUCK OFF
BAD GOYIM
SUPPORT INTEL, FEMINIST FREQUENCY AND JIZZ TIM

Not being a kike jew company gets you fans woah

Zen2's die is probably thinner (like RR), so I'm not sure if it will come soldered. BTW this is the same reason why Intel cannot solder their IHS, it'd break the die once the solder cools down.

> it'd break the die once the solder cools down.

If on LN2, sure. Without active cooling solder won't crack the die (in b4 Der8aur - he was talking about extreme temperature swings like those associated with LN2 cooling).

techpowerup.com/241593/amd-ryzen-2000-series-pinnacle-ridge-cpus-get-soldered-ihs
They've said it'll be soldered unlike Raven Ridge, so we should be gucci.

Zen+

>but isn't that what intel has been doing
This is not an argument. Both vendors doing the same bad thing does not make it alright. It's like AMD using toothpaste instead of solder on their new APUs. That's not alright. It's not acceptable. It's a total outrage! Yes, yes, Intel's using toothpaste on all their CPUs, even the high-end. This doesn't make it alright. It's still fucking toothpaste instead of solder under the lid.

>yfw amd is making kaby lake look like a good skylake upgrade
You must be a moron

>It's like AMD using toothpaste instead of solder on their new APUs.
AMD always used TIM in APUs.

the kaveri refresh aka "Godavari" has soldered IHS

> PDF render acceleration, go!

the fuck?

>that isnt fair cpus arent allowed to improve process and ipc at the same time

Zen2 7nm suicide watch countdown when?

That particular test was probably choking somewhere in the cache so AMD just fixed it, it's not uncommon for new architectures to simply shit themselves on certain functions and only get fixed in the line refresh

tl;dr pipeline stall/bubble
en.wikipedia.org/wiki/Bubble_(computing)

AMD has a bunch of these, that's why you see the IPC difference between AMD and Intel is anywhere from -5% to 20% depending on application.
Intel has been rehashing their core for years so they've mostly fixed these issues.

300 milliherz? I agree, that wouldn't be much of an upgrade.
Luckily, it's 300 MHz, so it's a substantial improvement.

> actually thinks pipeline bubbles are an issue with modern speculative/OOO cores

unless you can actually point me to the spot in Agner's tables etc. of some unnecessarily long latency instructions of Zen, I will call you an idiot.

memory stalls, which are not at all the same thing despite superficial similarities, are another issue altogether, and Ryzen 1 is well know to have been pretty lax with latencies outside the CCXs in the name of power saving and supporting large NUMA node counts.

300MHz under the same TDP, question is fmax.

Also it's the 2700X, not 2800X

If you seriously think "pdf rendering" chokes the memory controller in any measurable way, you're an idiot.
Nothing else but a borked function or, total cache or pipeline flush can explain the large increase in this microbenchmark.

jesus christ, the comment above a call for speculation on exactly what behavior that particular benchmark was stumbling on.

the huge boost being particular to the multithreaded variant suggests that it's not a proper fix for a microcode hackjob in Zeppelin, unless it's something specific to some form of IPC primitive handling. the more interesting alternative is that AMD tweaked intra-CCX IF fast paths for better coherency latencies (at cost of bandwidth, more power, or who knows).

the initial comment on bubbles is still retarded, as is your assumption that anyone could sincerely believe that any CPU would add in app-specific acceleration that somehow didn't require an instruction set change.

this processor is an engineering sample, final silicon will clock higher just like RR

2200G: 3000/3500 ES -> 3500/3700 final
2400G: 3300/3500 ES -> 3600/3900 final

op is a faggot and a jew

Unlikely.

It'll clock higher overclocked, probably 4.4 or 4.5, but these seem likely for their stock clocks for that model.

show us 2600x and 2800x

No it's not, it's a Qualification Sample that don't really differ from the final consumer product usually.

Regardless of that, I don't get OP. It was always clear that this is merely a refresh (hence the Zen+ naming not Zen 2), and as such it's a greater improvement than any of refreshes intel has EVER done.

Wonder if I should just get an 1800x now or not. Running a Phenom 1090t II @ 3.2Ghz.

it's better to wait 1 month and getting something a bit better with a newer motherboard

What do you use it for though. I too still have a Phenom II and I can't really justify the spending for my usage.

Intel was able to do that sort of improvement without even changing lithography or process with Haswell. The launch Haswells were 200-400MHz behind the Haswell Refresh that came out a year later.

>I've said that the AM4 "upgrade path" "selling point" bullshit shills here has been silly and stupid numerous times before. It should be more obvious now that the chips are here. You don't buy a $300 1700x just to "upgrade" to a max 5% faster $300 2700x a year later.

hothardware.com/news/amd-confirms-am4-socket-support-future-ryzen-processors-2020